Closed iory closed 7 years ago
2016-12-09 19:09 GMT+09:00 Masaki Murooka notifications@github.com:
default should be "/image_publisher/output" to subscribe image_publisher topic in default.
What does this mean? also look at https://github.com/ros-perception/opencv_apps/blob/indigo/src/node/standalone_nodelet_exec.cpp.in#L41
-- ◉ Kei Okada
What does this mean?
In this launch file, image_publisher node is run: https://github.com/ros-perception/opencv_apps/pull/48/files#diff-e102ca2ed99024189fade3737992c127R16 So, I thought that this launch file intends to connect from image_publisher to color_filter.
I talked with @iory offline, and he said that this is not sample file so image_publisher should be removed from this launch file.
could you update images at -> https://github.com/ros-perception/opencv_apps/pull/48#issue-184543025 so that everyone can understand the benefit of this node?
need better one, which we can use in wiki
-- ◉ Kei Okada
2016-12-14 12:37 GMT+09:00 iory notifications@github.com:
[image: image] https://cloud.githubusercontent.com/assets/4690682/21168775/212a77ca-c1fa-11e6-9b5c-139cff542666.png
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ros-perception/opencv_apps/pull/48#issuecomment-266934059, or mute the thread https://github.com/notifications/unsubscribe-auth/AAeG3IO8ZHBRSGecI9iX4suUbkxti57Qks5rH2SXgaJpZM4KdeWt .
Images of my use case.
sorry i think original input is ok, because every one can get that rosbag image, I just want to understand that what who created sample image want to extract. Human face? cloth? floor????
-- ◉ Kei Okada
2016-12-14 14:23 GMT+09:00 Masaki Murooka notifications@github.com:
Images of my use case. [image: before] https://cloud.githubusercontent.com/assets/6636600/21170488/cece7918-c208-11e6-8d6a-7909c607dd90.png [image: after] https://cloud.githubusercontent.com/assets/6636600/21170489/cfb41716-c208-11e6-97d7-8b63c796481d.png
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ros-perception/opencv_apps/pull/48#issuecomment-266945744, or mute the thread https://github.com/notifications/unsubscribe-auth/AAeG3OZn4EqP9l1TZJ9xoGJMHT6yX9Hlks5rH31cgaJpZM4KdeWt .
human's blue cloth.
Sample program's parameter updated.
hsi_color_filter
rgb_color_filter
@iory , can I see original image? It is difficult to imagine where human's blue cloth is? Are target regions of filtering same for the examples of hsi_color_filter and rgb_color_filter? They are quite different.
Original image is here.
It is difficult for me to separate colors clearly in RGB color spaces. I think this is advantage of HSI color filter over RGB color filter.
I see. @k-okada , are images in https://github.com/ros-perception/opencv_apps/pull/48#issuecomment-267241077 good enough for wiki document? Or do you have any good idea to make good sample images? (now cloth is target for filtering. should we filter other region?)
Can we show only human skins?
2016年12月20日(火) 10:38 Masaki Murooka notifications@github.com:
I see. @k-okada https://github.com/k-okada , are images in #48 (comment) https://github.com/ros-perception/opencv_apps/pull/48#issuecomment-267241077 good enough for wiki document? Or do you have any good idea to make good sample images? (now cloth is target for filtering. should we filter other region?)
— You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub https://github.com/ros-perception/opencv_apps/pull/48#issuecomment-268129712, or mute the thread https://github.com/notifications/unsubscribe-auth/AAeG3CkdrrVSz-Kdgq2fCj5D4wPo-4NSks5rJzGjgaJpZM4KdeWt .
OK.
i agree this is very difficult to find good example (this means that we do not have good application for this node?), but we have to try as hard as possible.
range should be 0-256, since inRange() function lower <= src < uppder
, H shoulde be 0-360 and divide by 2?...
and you also need something like
if ( lower_color_range_[0] < upper_color_range_[0] ) {
cv::inRange(hsv_image, lower_color_range_, upper_color_range_,
output_image);
}else {
cv::Scalar lower_color_range_0 = cv::Scalar( 0, s_min_, i_min_, 0);
cv::Scalar upper_color_range_0 = cv::Scalar(h_max_/2, s_max_, i_max_, 0);
cv::Scalar lower_color_range_360 = cv::Scalar(h_min_/2, s_min_, i_min_, 0);
cv::Scalar upper_color_range_360 = cv::Scalar( 360/2, s_max_, i_max_, 0);
cv::Mat output_image_0, output_image_360;
cv::inRange(hsv_image, lower_color_range_0, upper_color_range_0, output_image_0);
cv::inRange(hsv_image, lower_color_range_360, upper_color_range_360, output_image_360);
output_image = output_image_0 | output_image_360;
}
Thanks to the k-okada's PR (https://github.com/iory/opencv_apps/pull/1), hsv and hls color filter were added into color_filter.
hsv_color filter
hls_color filter
Cloud you merge this?
@iory @mmurooka please update doc at http://wiki.ros.org/opencv_apps
note that do not forget to add New feature as of version 1.11.15
OK
Updated docs
Thank you!
thank you
-- ◉ Kei Okada
2017-03-07 18:32 GMT+09:00 iory notifications@github.com:
Updated docs
— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub https://github.com/ros-perception/opencv_apps/pull/48#issuecomment-284668865, or mute the thread https://github.com/notifications/unsubscribe-auth/AAeG3Fw1KxCFuhnqcU2SScGE6CwXC9cUks5rjSQwgaJpZM4KdeWt .
add rgb_color_filter![rgb_color_filter](https://cloud.githubusercontent.com/assets/4690682/19607921/111feb10-9808-11e6-8295-a5752eac1035.png)
and hsi_color_filter